***********************************************************************************************************
**REPLICATION INSTRUCTIONS 

"How does the educational cleavage stack up against the classic against the classic cleavages of the past?"

Liesbet Hooghe and Gary Marks
West European Politics (accepted Dec 17, 2024)

For questions, please contact hooghe@unc.edu 

************************************************************************************************************



"MASTER_CLEAVAGE dec2024.do" is the single do file for replicating the figures and tables in the paper. The datasets below have been prepared to contain the variables that you need for replication. The do file contains information on where to obtain the original datasets used to generate these replication data (see also Online Appendix A (Data Sources).

In order of use: 

1. Trends in P for education, occupation, and gender (Figures 1-3)

Use "EBESSnonvotersmerge_dec24.dta". This dataset contains P values on education, occupation, and gender for each political party and party bloc in eight European countries since 1973 (or 1975 for occupation). This pools information from the Eurobarometer time series and the European Social Survey time series. 

For operationalization, see Online Appendix B (Coding TAN, GAL, LEFT, and RIGHT parties) and Online Appendix C (Operationalization of variables by data source)



2. The classical cleavage benchmarks (Tables 1-3) 

For operationalization, see Online Appendix C (Operationalization of variables by data source)

For instructions on how to extract each of the components of P from the crosstabulation between occupation and party vote choice, please see online Appendix D (Calculation of PCI and its components -- example 1).

2.1.  German SPD 1957 and 2017

Use "Germany 1957 working.dta" for the earlier time point
Use "ESS_master_12_18_2024.dta" for the later time point

2.2.  Norwegian Labor Party 1957 and 2020

Use "norway1957working.dta"
Use "ESS_master_12_18_2024.dta"

2.3. British Labour Party 1964 and 2020

Use "BNES working.dta"
Use "ESS_master_12_18_2024.dta"



3. GAL and TAN P in time (Tables 4-5) 

For operationalization, see Online Appendix C (Operationalization of variables by data source). 

For instructions on how to extract each of the P components from a crosstabulation between the social characteristic of interest and party vote choice, and how to aggregate this information up across years and countries, please see online Appendix D (Calculation of PCI and its components - example 2)  

3.1. Older time period (1990-94 for GAL, 1995-2009 for TAN)

Use "EBtrend_dec2024withnonvoters.dta"

3.2. Recent time period (2016-2020)

Use "ESS_master_12_18_2024.dta"


You can find the P components by country and time period in the following excel files: 
*for GAL parties:  "GAL P breakdown using survey data closed system dec 2024.xlsx"
*for TAN parties:  "TAN P breakdown using survey data closed system dec 2024.xlsx"


4. Cleavaged America (Figure 4 and Table 6)

The empirical basis is ANES. The information needed for Figure 4, Table 6 and discussion in the text is prepared in three separate datasets. 

4.1. Lower-educated Whites -- P over time (Figure 4)

Use "ANES Pvalues dec 2024.dta"

4.2. Lower-educated Whites -- P and its components (Table 6)

Use "ANES working dec2024.dta"

4.3. Recent trends (2022)

Use "anes_pilot_2022_data_20221214.dta" 